Coffee Secrets in the Spectrum Spotlight: New Arabica Diterpenoids with Anti-Diabetic Potential Found

Scientists from the Chinese Academy of Sciences have shown that roasted Coffea arabica beans contain previously undescribed diterpene esters that inhibit the enzyme α-glucosidase, a key accelerator of carbohydrate absorption. The team combined “fast” ¹H-NMR fraction imaging and LC-MS/MS with molecular networking to first map out the most “bioactive” zones in the extract and then extract specific molecules from them. As a result, three new compounds with moderate α-glucosidase inhibition were isolated and three more related “trace” candidates were identified by mass spectra.

Background of the study

Coffee is one of the most chemically complex food matrices: the roasted grain and the drink simultaneously contain hundreds to thousands of low-molecular compounds - from phenolic acids and melanoidins to lipophilic diterpenes of coffee oil. It is the diterpenes (primarily derivatives of cafestol and kahweol) that attract special attention: they are associated with both metabolic effects (including the influence on carbohydrate metabolism) and cardiac markers. An important detail is that in the grain they exist almost entirely in the form of esters with fatty acids, which increases hydrophobicity, affects extraction during brewing and potential bioavailability in the body.

From the point of view of preventing postprandial hyperglycemia, a rational target is the enzymes that break down carbohydrates in the intestine, primarily α-glucosidase. Inhibitors of this enzyme (mechanically similar to the "pharmaceutical class" of acarbose/voglibose) slow down the breakdown of disaccharides and reduce the rate of glucose entry into the blood. If among the natural components of coffee there are substances with moderate activity against α-glucosidase, they can potentially "soften" the sugar peaks after meals and complement dietary strategies for glycemic control - of course, provided that they are in sufficient concentrations in real food and have confirmed bioavailability.

The classic problem of natural sources is looking for a needle in a haystack: active molecules are often hidden in the "tail" fractions and are present in trace amounts. Therefore, bioactivity-oriented dereplication is increasingly used: first, a "portrait" of fractions is taken using fast NMR, they are tested in parallel for the target enzyme, and only then the "hot" components are specifically caught using high-performance chromatography. The approach is complemented by molecular networking LC-MS/MS, which groups compounds related by fragmentation and allows one to notice rare analogs even without complete isolation. Such an analytical tandem accelerates the path from "there is an effect in the fraction" to "here are specific structures and their family."

Finally, the technological and nutritional context. The profile and amount of coffee diterpenes depend on the variety (Arabica/Robusta), the degree and mode of roasting, the extraction method (oil/water media) and the filtration of the drink. To translate laboratory findings into practice, it is necessary to understand in which products and with which methods of preparation the required levels of compounds are achieved, how they are metabolized (hydrolysis of esters, conversion to active alcohol forms) and whether they conflict with other effects. Hence the interest in works that do not simply “take spectra”, but purposefully search for new coffee diterpenoids with a validated biological target - a step towards substantiated functional ingredients, and not towards another “myth about the benefits of coffee”.

What was done (and how this approach differs)

• The roasted Arabica extract was divided into dozens of fractions and their “portraits” were assessed using ¹H-NMR, while simultaneously measuring the inhibition of α-glucosidase for each fraction. On the heat map, the active zones immediately “floated” to the top.
• The “hottest” fractions were purified by HPLC, isolating three main peaks (tR ≈ 16, 24 and 31 min; UVmax ~218 and 265 nm) - these turned out to be new diterpenoid esters (1-3).
• In order not to lose rare related molecules, a molecular LC-MS/MS network was constructed: three more “trace” analogues (4–6) were found from fragment clusters, which could not be isolated, but were confidently recognized by the MS signature.

What was found - in essence

• Three new diterpenoid esters (1-3) from Arabica showed moderate activity against α-glucosidase (in the micromolar range of IC₅₀; n=3). This is an important “mechanistic” signal for carbohydrate metabolism.
• Three more analogues (4-6) were mapped by HRESIMS/MS and shared fragments m/z 313, 295, 277, 267 - a typical "family" signature for coffee diterpenes. The formulas were confirmed by HRMS (e.g. C₃₆H₅₆O₅ for compound 1).
• Context: Coffee diterpenes (primarily cafestol and kahweol derivatives) in coffee are almost entirely (≈99.6%) present as fatty acid esters in the coffee oil; they are usually present in higher amounts in Arabica than in Robusta.

Why is this important?

• Functional coffee ≠ only caffeine. Diterpenes have long been “suspected” of antidiabetic and antitumor effects; cafestol already has in vivo and in vitro data on stimulating insulin secretion and improving glucose utilization. New esters expand the chemical family and provide fresh “hooks” for nutraceuticals.
• Methodology accelerates discoveries. The combination of ¹H-NMR "broad stroke" + LC-MS/MS-networking allows to quickly de-replicate known molecules and focus on new ones, saving months of routine.

Coffee under the microscope: what exactly was measured

• Heat map of ¹H-NMR fractions with superimposed α-glucosidase activity (IR, 50 μg/ml) → highlighting the “top fraction”.
• Structural elucidation 1-3: full 1D/2D NMR + HRMS set; key correlations (COSY/HSQC/HMBC) are shown.
• Molecular network (MN-1) for "neighbor search" 4-6; nodes 1-3 are located next to each other - additional confirmation of "one chemical family".

What does "in the kitchen" mean (careful while the lab is running)

• Coffee is not only a source of energy, but also biomolecules that potentially moderate glycemic peaks (via α-glucosidase). But extrapolation is limited: the activity was measured in enzyme and cell assays, not in clinical RCTs.
• The path to a "functional ingredient" is standardization, safety, pharmacokinetics, and human evidence. For now, it is correct to talk about chemical candidates, not "medicinal coffee."

Details for the curious

• UV profile of new esters: 218 ± 5 and 265 ± 5 nm; HPLC retention ~16/24/31 min.
• HRMS formulas (M+H)⁺: e.g. C₃₆H₅₆O₅ (1), C₃₈H₆₀O₅ (2), C₄₀H₆₄O₅ (3); for 4-6 - C₃₇H₅₈O₅, C₃₈H₅₈O₅, C₃₉H₆₂O₅.
• Where in the beans are these substances? Mostly in coffee oil, esteroforms with palmitic/linoleic acids predominate.

Limitations and what's next

• In vitro ≠ clinical effect: α-glucosidase inhibition is only a marker test. Bioavailability, metabolism, animal models and then RCTs in humans are needed.
• Roasting changes the chemistry. The composition and proportions of diterpenes depend on the variety, thermal regime and extraction - for real products, technological optimization will be required.
• The tool itself is universal. The same "NMR + molecular network" can be directed at tea, cocoa, spices - anywhere where there are complex extracts and the hunt for microcomponents.

Conclusion

The researchers "illuminated" Arabica with two devices at once and extracted six new diterpene esters from the coffee oil, three of which were isolated and confirmed to be active against α-glucosidase. This is not yet a "coffee pill", but a convincing chemical trace to functional ingredients for controlling carbohydrate metabolism - and a clear example of how smart analytical approaches are accelerating the hunt for molecules of benefit in our usual products.

Source: Hu G. et al. Bioactive oriented discovery of diterpenoids in Coffea arabica basing on 1D NMR and LC-MS/MS molecular network. Beverage Plant Research (2025), 5: e004. DOI: 10.48130/bpr-0024-0035.